【作者】 杨族桥；

【导师】 郭庆胜；

【作者基本信息】 武汉大学 ， 地图学与地理信息系统， 2009， 博士

【摘要】 尺度是指在研究某一物体或现象时所采用的空间或时间单位,又可指某一现象或过程在空间和时间上所涉及到的范围和发生的频率,简单地说,尺度就是客体在其“容器”中规模相对大小的描述。尺度问题是地理信息科学以及地图学、遥感科学、地理学、水文学等学科中必须研究的核心理论问题。地理信息、地理现象以及地理过程在认知、获取、建模、测度、解译、可视化、存储、传输等处理过程中的任何一个环节都脱离不了对尺度的讨论,随着地理信息科学的应用领域的不断扩大和需求层次的日益提高,特别是数字化环境下,一方面,获取的空间数据量非常巨大,另一方面,人们对于随时随地的使用高精度的数据有着更高的期望,这就要求数字地形数据能够以高效的方式在各种介质上传递并再现。另外,对于如此海量的数据无论使用何种逼真的可视化技术,也难以让人迅速有效的挖掘其中需要的知识,它已经远远超出了人脑认知的尺度范围。因此,利用多尺度理论和技术多层次观察、处理、分析地理现象、描述地理过程变得越来越重要。空间数据的多尺度处理与表达,已成为建立多空间尺度数据库,实现多尺度GIS或无级比例尺GIS的核心内容,同时建立尺度依赖的空间数据表达模型,实现地理信息多尺度自动综合是解决空间数据多尺度处理与表达的关键技术。尺度和空间尺度理论是地理学、地图学和地理信息科学等学科中研究的热点问题,而且这些理论有着深厚的空间认知、信息科学、计算数学、计算机图形学以及工程应用的基础。本文在仔细探讨地理信息的空间尺度理论基础上,研究了导致多尺度现象产生的机理,总结了现有的对空间信息进行多尺度、多分辨率分析的技术和方法,在深入分析DEM数据多尺度表达的内涵基础上,结合地形特征线的重要作用和渐进式地图综合的基本思想,利用小波多分辨率分析原理和尺度理论的最新成果,提出了一个对DEM进行渐进式多尺度表达的理论框架,实现对DEM的渐进式多尺度表达和自动综合。本论文主要介绍和研究了以下几个问题：一)从地理信息科学的学科发展、理论研究以及社会需求等方面阐述空间数据多尺度处理问题研究的科学意义和必要性,叙述本论文的研究目标、主要研究内容以及论文章节的组织结构。二)介绍了当今空间数据多尺度处理与表达研究领域的研究主题和国内外研究发展的状态。系统地分析和归纳空间尺度的相关研究成果与存在问题以及DEM数字地形分析中的尺度类型和尺度变换方法,为进一步的开展研究提供目标和技术方向。三)分析了当前空间数据库存在的不足,研究了渐进式地图综合的基本思想及其形式化描述,指出渐进式思想是建立多尺度纵向连接关系的有效方法,结合地理空间的尺度特征,提出基于渐进式综合思想的多尺度表达模型研究了渐进式地图综合的思想,以使DEM数据处理过程中具有尺度维的分析处理和表达能力。四)研究了地形特征线的推理。首先基于约束Delaunay三角网(简称CDT)建立了等高线的层次结构,并利用这种关系实现地形分类以及等高线走向的调整。其次提取CDT中的平三角形区域,对于每一个平区域提取概略性地形特征线,对地形特征线上每一个点的高程值按一定的算法进行加权内插,得到更符合实际地形的特征线,将地形特征线加入TIN修正平三角形区域。再次研究了一种从成组等高线中提取地形特征线的方法：1)构造等高线的约束Delaunary三角网；2)计算等高线上各顶点的转角、曲率值及形状指数,并依据给定的阈值确定候选地形特征点集；3)按照相应的原则将符合条件的点连接成地形结构线；4)并对其进行优化。接下来还研究基于等高线约束的Delaunary三角网基础上求算每个Delaunary三角形(面片)的流向,并根据各种流向的组合关系来对Delaunary三角形的公共边进行分类,将相关的边连接得出地形结构线的算法。五)研究了基于小波分析理论的空间数据多尺度处理模型。首先简述小波分析理论的基本原理,分析空间数据多尺度处理与表示研究中应用小波分析理论的可能性和必要性；然后从小波分析理论的应用角度,研究空间数据多尺度处理与表示的原理和模型。基于小波分析理论,研究并提出满足尺度依赖空间数据模型要求并与之相配套的空间数据多尺度处理模型,完成基于二进制、多进制的多尺度表达模型的实现以及实验结果的分析,并研究基于提升变换的多尺度表达模型和实现原理,从性能上与前面的方法进行了比较。六)在改进的地形特征线的提取算法的基础上,实现了基于多尺度DEM的地形特征线的提取,并利用结构线的重叠程度来验证本文提出的方法的合理性。最后对本论文研究作全面总结,并指出下一步的研究方向和问题。更多还原

【Abstract】 Scale is the space or time unit used for study of a geographical object or a geographical phenomenon. It also refers to the scope and the frequency of the phenomenon or process involved in space or time domain. In brief, it is the description of the relative size of the object in its container. Scale is the foundational issue that should be studied in geographic information science and cartography, remote sensing science, geography, hydrology and other disciplines.In any aspect of the process of cognition, acquisition, modeling, measurement, visualization, storage, and transmission of geographic information, phenomenon and process, scale is inseparable from the discussion. As the application field of geographic information science is expanding and the society requirements is increasing. Especially under the digital environment, on the one hand, the spatial data acquired is huge; on the other hand, people have high expectation for use of high-precision data at anytime and anywhere. This requires the digital terrain data to be able to efficiently transmit among a variety of media and be represented. In addition, for such a mass of data regardless of what kind of realistic visualization technology, it will be difficult for people to quickly and effectively mine the required knowledge. It has gone far beyond the scale scope of human brain cognitive. Therefore, using multi-scale theory and technology to observe, process, and analyze the geographic phenomenon in multiple levels, and describe geographic process become so important. The multi-scale processing and representation of spatial data has become the fundamental content in creating multi-space scale database and realizing multi-scale GIS or scale-free GIS or on the fly GIS, at the same time, building spatial data representation model dependent on scale and realizing multi-scale automatic generalization of geographic information is the key technology for processing and representing multi-scale spatial data.Scale and spatial scale theory is a hot research issue in geography, cartography and geographic information science, and these theories are on the basis of spatial cognition, information science, computational mathematics, computer graphics, as well as application.Based on the spatial scale theory of geographic information, this dissertation studies the mechanism of multi-scale phenomenon; summarizes present multi-scale and multi-resolution technologies and methods of spatial information. After analysing the connotation of DEM multi-scale representation in-depth, merging the important function of terrain characteristic lines and the basic idea of progressive map generalization, making use of wavelet multi-resolution analysis theory and the newest results of scale theory, a theoretical frame of progressive multi-scale representation for DEM is proposed and progressive multi-scale representation and automatic generalization of DEM is accomplished.The dissertation includes the following sections:1) The scientific significance and necessity of multi-scale processing of spatial data is discussed from the aspect of academic development and theoretical research of geographic information and social demands. The research general goal, main research content and the organization of dissertation is given.2) Related works. Analyzes the research content and the present results of spatial scale, introduces the researching topics and development status at home and abroad in the area of spatial data multi-scale processing and representation. It systematically analyzes and summarizes relevant researching results and existing problems of spatial scale, scale type and scale transform methods in DEM digital terrain analysis. It mainly analyzes scale characteristics of geographic space to provide goal and technical direction for further research.3) Research on the multi-scale representation model based on progressive generalization. It analyzes the inefficiency of current spatial database, studies the basic idea and its formal description of progressive map generalization, and points out that progressive idea is an effective method to set up multi-scale vertical connection relationship. Combining with the scale characteristics of geographic space, it puts forward and studies the multi-scale representation model based on progressive generalization so that there is analyzing, processing and expressing ability of scale-dimension when dealing with DEM data.4) Study on the reasoning of terrain characteristic lines. First of all, it sets up constrained Delaunay triangulation (CDT) and a hierarchy structure of contour lines, and realizes terrain classification and adjustment of the trend of contours. Then it extracts the flat triangle regions from CDT, and extracts general terrain characteristic lines from each flat region. Conduct weighting interpolation on the elevation value of each point in the terrain characteristic lines according to certain algorithms and achieve characteristic lines more in line with the actual terrain. Add the terrain characteristic lines into TIN to amend the flat triangle region. And then the method of extracting terrain characteristic lines from group contours is proposed:Firstly, Constrained Delaunay Triangulation of contours is constructed; secondly, the turn angle, curvature and shape index values of each vertex on the contour are calculated, and the candidate terrain characteristic points are classified by the given threshold; thirdly, connect the points matching conditions into terrain characteristic lines according to corresponding principles; fourthly, the line is optimized. Compute the flow direction of each triangle in Constrained Delaunay Triangulation based on contours, and classify the public sides of Constrained Delaunay triangle according to the combination relationship between various flow directions. Finally, the algorithm connecting the related sides to obtain terrain characteristic lines is proposed in dissertation.5) The multi-scale processing model of spatial data based on wavelet analysis theory. It briefly introduces the basic principle of wavelet analysis theory, analyzes the possibility and necessity of using wavelet analysis theory in processing and representing multi-scale spatial data. It analyzes the principle and model of multi-scale processing and representation of spatial data from the application aspect of wavelet analysis theory. Based on wavelet analysis theory, it proposes multi-scale processing model of spatial data that meets the requirements of and matches scale-dependent spatial data model. It realizes multi-scale representation model based on binary and multi numbering system, and makes experimental analysis. It also studies the multi-scale representation model and implementation principle based on lifting transform, and does comparison with former methods.6) Based on the improved extracting algorithm of terrain characteristic lines, terrain characteristic lines based on multi-scale DEM are extracted and the reasonableness of the method proposed in this dissertation is verified by the overlap rate of terrain characteristic lines.At last, a comprehensive conclusion is summed up and the research directions and problems to be solved in future are pointed out in the paper.更多还原